This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [RFA] c++/11734 revisited
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: Keith Seitz <keiths at redhat dot com>
- Cc: Tom Tromey <tromey at redhat dot com>, gdb-patches at sourceware dot org
- Date: Wed, 26 Jan 2011 23:35:26 +0100
- Subject: Re: [RFA] c++/11734 revisited
- References: <4D002672.5090504@redhat.com> <m3oc8u1uzs.fsf@fleche.redhat.com> <4D07CD68.5000700@redhat.com> <20110124173745.GA31199@host1.dyn.jankratochvil.net>
On Mon, 24 Jan 2011 18:37:45 +0100, Jan Kratochvil wrote:
> > Revised patch attached.
>
> As a summary:
>
> "physname" patch I call:
> 42284fdf9d8cdb20c8e833bdbdb2b56977fea525
> http://sourceware.org/ml/gdb-cvs/2010-03/msg00082.html
> dwarf2_physname patchset:
> [RFA] dwarf2_physname FINAL
> http://sourceware.org/ml/gdb-patches/2010-03/msg00220.html
>
> Original pre-physname code for pr11734::foo() did not call decode_compound but
> called decode_variable which resolved it from the fully qualified name.
>
> post-physname code tried to resolve pr11734::foo() via decode_compound which
> did not support such case and it fails now on HEAD.
>
> Alternatively to your patch the current HEAD can be modified to call
> decode_variable again as in pre-physname era which resolves it correctly but
> for some reason you prefer to resolve pr11734::foo() via decode_compound
> instead which makes sense.
I have to withdraw my preliminary approval. Together with
[RFA] c++/12273
http://sourceware.org/ml/gdb-patches/2010-12/msg00264.html
both patches duplicate the decode_variable functionality by the new
implementation in decode_compound.
But this duplication is still not complete, for example it does not handle
properly namespaces. It had only a luck the existing namespace tests
(nsusing.exp) do not exploit it.
I have attached such testcase on top of your this patch:
Re: [RFA] c++/11734 revisited
http://sourceware.org/ml/gdb-patches/2010-12/msg00263.html
with results:
PASS: gdb.cp/pr11734.exp: break func
PASS: gdb.cp/pr11734.exp: break 'func'
^^^^ = test like nsusing.exp
FAIL: gdb.cp/pr11734.exp: break pr11734::foo() (got interactive prompt)
FAIL: gdb.cp/pr11734.exp: break 'pr11734::foo()' (got interactive prompt)
FAIL: gdb.cp/pr11734.exp: break pr11734::foo(char*) (got interactive prompt)
FAIL: gdb.cp/pr11734.exp: break 'pr11734::foo(char*)' (got interactive prompt)
FAIL: gdb.cp/pr11734.exp: break pr11734::foo(int) (got interactive prompt)
FAIL: gdb.cp/pr11734.exp: break 'pr11734::foo(int)' (got interactive prompt)
^^^^ = these funcs get called by the C++ inferior but GDB cannot resolve them.
decode_compound itself should comply to namespacing, just dropping to
decode_variable as before is not enough anyway as decode_variable cannot
resolve class fields which have no global symbols/minsym.
The new plan of G++ parser/plugin is probably not considerable now as a fix.
With the very hacky drop into decode_variable the testcase works but this
patch of mine is sure not usable.
Thanks,
Jan
--- a/gdb/linespec.c
+++ b/gdb/linespec.c
@@ -783,7 +783,7 @@ decode_line_1 (char **argptr, int funfirstline, struct symtab *default_symtab,
/* Does it look like there actually were two parts? */
- if (p[0] == ':' || p[0] == '.')
+if (0) // if (p[0] == ':' || p[0] == '.')
{
/* Is it a C++ or Java compound data structure?
The check on p[1] == ':' is capturing the case of "::",
@@ -868,7 +868,22 @@ decode_line_1 (char **argptr, int funfirstline, struct symtab *default_symtab,
}
else
{
- p = skip_quoted (*argptr);
+ const char *lang_breakchars;
+ char *breakchars, *s, *end;
+
+ lang_breakchars = current_language->la_word_break_characters ();
+
+ breakchars = alloca (strlen (lang_breakchars) + 1);
+ strcpy (breakchars, lang_breakchars);
+ end = &breakchars[strlen (breakchars)];
+
+ s = strchr (breakchars, ':');
+ if (s != NULL)
+ {
+ *s = *--end;
+ *end = 0;
+ }
+ p = skip_quoted_chars (*argptr, NULL, breakchars);
}
/* Keep any template parameters. */
--- a/gdb/testsuite/gdb.cp/pr11734-1.cc
+++ b/gdb/testsuite/gdb.cp/pr11734-1.cc
@@ -20,11 +20,21 @@
#include "pr11734.h"
+namespace A { void func () {} }
+
+using namespace A;
+
int
main ()
{
pr11734 *p = new pr11734;
p->foo ();
+
+ func ();
+
+ pr11734::foo ();
+ pr11734::foo ((char *) "");
+ pr11734::foo (0);
+
return 0;
}
-
--- a/gdb/testsuite/gdb.cp/pr11734-2.cc
+++ b/gdb/testsuite/gdb.cp/pr11734-2.cc
@@ -20,8 +20,11 @@
#include "pr11734.h"
+namespace A {
+
void
pr11734::foo(void)
{
}
+}
--- a/gdb/testsuite/gdb.cp/pr11734-3.cc
+++ b/gdb/testsuite/gdb.cp/pr11734-3.cc
@@ -20,8 +20,11 @@
#include "pr11734.h"
+namespace A {
+
void
pr11734::foo (int a)
{
}
+}
--- a/gdb/testsuite/gdb.cp/pr11734-4.cc
+++ b/gdb/testsuite/gdb.cp/pr11734-4.cc
@@ -20,8 +20,11 @@
#include "pr11734.h"
+namespace A {
+
void
pr11734::foo (char *a)
{
}
+}
--- a/gdb/testsuite/gdb.cp/pr11734.exp
+++ b/gdb/testsuite/gdb.cp/pr11734.exp
@@ -34,6 +34,15 @@ if {![runto_main]} {
continue
}
+if [gdb_breakpoint "func"] {
+ pass "break func"
+}
+if [gdb_breakpoint "'func'"] {
+ pass "break 'func'"
+}
+
+gdb_test "ptype ${class}::l" "type = long"
+
# An array holding the overload types for the method pr11734::foo. The
# first element is the overloaded method parameter. The second element
# is the expected source file number, e.g. "pr11734-?.cc".
--- a/gdb/testsuite/gdb.cp/pr11734.h
+++ b/gdb/testsuite/gdb.cp/pr11734.h
@@ -18,11 +18,15 @@
Please email any bugs, comments, and/or additions to this file to:
bug-gdb@gnu.org */
+namespace A {
+
class pr11734
{
public:
- void foo ();
- void foo (int);
- void foo (char *);
+ static void foo ();
+ static void foo (int);
+ static void foo (char *);
+ long l;
};
+}